Elizabeth Phillips

This is one of the largest long-term studies following an SJS/TEN reaction that reports long-term survey-based sequelae and adds to the understanding of SJS/TEN as a condition with chronic complications long beyond the acute reaction. We report significant long-term life impacts among survivors of SJS/TEN in the USA. Additional follow-up and long-term care of this patient population is of the utmost importance.

Immune checkpoint inhibitors (ICIs) are paradigm-shifting cancer treatments that are increasingly associated with Stevens-Johnson syndrome, toxic epidermal necrolysis (SJS/TEN) and other life-threatening cutaneous reactions. Differentiating ICI-induced true SJS/TEN from SJS/TEN-like reactions is difficult, the latter of which may be distinct lichenoid or bullous reactions.1-3 In some cases, ICI-related SJS/TEN-like reactions occur in association with human leukocyte antigen (HLA)–restricted drug culprits like allopurinol, suggesting a 2-hit mechanism.4 With increasing ICI use, a clearer understanding of their role in SJS/TEN is critical.

Methods
We analyzed 13 986 839 deduplicated Food and Drug Administration Adverse Event Reporting System (FAERS) reports (2013-2023), containing 17 495 patients with SJS/TEN. We assessed the impact of ICI using logistic regressions adjusted for age, sex, cancer, polypharmacy, and strong (lamotrigine, trimethoprim-sulfamethoxazole, phenytoin, allopurinol, carbamazepine) or weak (azithromycin, clarithromycin, erythromycin, ciprofloxacin, levofloxacin, moxifloxacin, and acyclovir) culprit exposure.

To assess latency patterns, we performed Cox proportional hazards analyses among patients with SJS/TEN with documented latency. In 1 model, we compared latency between ICI-attributed and non–ICI-attributed cases, classifying primary suspect (PS) by ICI mechanism. In another, we used time-dependent Cox regression with interval splitting to dynamically update exposure to programmed cell death 1 (PD-1) and its ligand (PD-L1), cytotoxic T-lymphocyte antigen 4 (CTLA-4), or lymphocyte-activation gene 3 (LAG-3) inhibitors. Both models incorporated the same covariates as the logistic regression. Additional eMethods are in Supplement 1.

Results
In a multivariable logistic regression (Table), ICI exposure was associated with increased risk of SJS/TEN (adjusted odds ratio [aOR], 9.14; 95% CI, 8.42-9.93; P < .001). Strong culprit drugs were the strongest independent predictors of SJS/TEN (aOR, 14.31; 95% CI, 13.77-14.87). Importantly, cancer diagnosis was inversely associated with SJS/TEN risk (aOR, 0.60; 95% CI, 0.58-0.63). Interaction terms revealed additive synergy between ICI exposure and culprit drugs. The ICI–strong culprit interaction yielded an attributable proportion (AP) of 0.38, indicating that 38% of the risk in co-exposed patients was attributable to interaction. For ICI-weak culprits, AP was even higher (0.52).